If you think of LLMs as being akin to lossy text compression of a set of text, where the compression artifacts happen to also result in grammatical-looking sentences, the question you eventually end up asking is "why is the compression lossy? What if we had the same thing but it returned text from its database without chewing it up first?" and then you realize that you've come full circle and reinvented search engines
200fifty
joined 2 years ago
"I know not with what technology GPT-6 will be built, but GPT-7 will be built with sticks and stones" -Albert Einstein probably
I think they were responding to the implication in self's original comment that LLMs were claiming to evaluate code in-model and that calling out to an external python evaluator is 'cheating.' But actually as far as I know it is pretty common for them to evaluate code using an external interpreter. So I think the response was warranted here.
That said, that fact honestly makes this vulnerability even funnier because it means they are basically just letting the user dump whatever code they want into eval() as long as it's laundered by the LLM first, which is like a high-school level mistake.
ok but for real... it's not great for finding actual answers to queries, but I find like 800x more interesting results with search.marginalia.nu than any other search engine. It's the only search engine that I find actively fun to just browse around on recreationally.
When I was a kid (Nat Nanny)[https://en.wikipedia.org/wiki/Net_Nanny] was totally and completely lame, but the whole millennial generation grew up to adore content moderation. A strange authoritarian impulse.
Me when the mods unfairly ban me from my favorite video game forum circa 2009
(source: first HN thread)
I like how the assumption seems to be that the thing users object to about "websites track your browsing history around the web in order to show you targeted ads" is... the "websites" part
Ew... stay away from my content, you creep!